nasa technical memorandum€¦ · ordinary end-fire operation the signals from microphones 1, 2, 3,...

NASA TECHNICAL NASA TM X- 62,331

MEMORANDUM

CD

x

(NASA-TM-X-6 23 3 1 ) A FOUR-ELEMENT END-FIRE 74-15913

l MICROPHONE ARRAY FOR ACOUSTIC MEASUREMENTS

IN WIND TUNNELS (NASA) 45 p BC $4.25z CSCL 09A UnclasG3/09 28340

A FOUR-ELEMENT END-FIRE MICROPHONE ARRAY FOR

ACOUSTIC MEASUREMENTS IN WIND TUNNELS

Paul T. Soderman and Stephen C. Noble

Ames Research Center

and 1S

U. S. Army Air Mobility R&D LaboratoryMoffett Field, Calif. 94035

January 1974

https://ntrs.nasa.gov/search.jsp?R=19740007800 2020-05-03T07:24:32+00:00Z

A FOUR-ELEMENT END-FIRE MICROPHONE ARRAY

FOR ACOUSTIC MEASUREMENTS IN WIND TUNNELS

Paul T. Soderman and Stephen C. Noble*

Ames Research Center

And

U. S. Army Air Mobility R&D Laboratory

Moffett Field, Calif. 94035

*Air Force Research Associate, now at Air Force Avionics Lab.,

Wright-Patterson Air Force Base, Ohio 45433

SUMMARY

A prototype four-element end-fire microphone array was designed and builtfor evaluation as a directional acoustic receiver for use in large windtunnels. The microphone signals were digitized, time delayed, summed, andreconverted to analog form in such a way as to create a directional responsewith the main lobe along the array axis. The measured array directivityagrees with theoretical predictions confirming the circuit design of theelectronic control module.

The array with 0.15 m (0.5 ft) microphone spacing rejected reverberationsand background noise in the Ames 40- by 80-Foot Wind Tunnel by 5 to 12 dB forfrequencies above 400 Hz.

NOTATION

c speed of sound

d .Microphone spacing, cm (ft)

dB decibel

f frequency of sound

U wind speed, m/s (f/s)

a directivity angle relative to array axis

X acoustic wavelength

INTRODUCTION

Various directional acoustic receivers are being developed for measuringnoise in the Ames 40- by 80-Foot Wind Tunnel. These receivers are needed todiscriminate between the desired acoustic signal and the reverberant andbackground noise in the wind tunnel. Ideally, sound close to that which wouldbe obtained in anechoic space could be measured by pointing a directionalreceiver having low sensitivity to wind across the wind tunnel at the source.This would alleviate masking of aircraft noise by a) reverberations due to aclosed hardwall test section, b) background noise from the wind tunnel fans,and c) microphone wind noise.

A linear end-fire* array of microphones is one such directional receiverbeing evaluated. Each microphone signal is delayed and summed with the othermicrophone signals so that sound waves traveling down the array axis sum inphase and sound waves off axis sum out of phase, tending to cancel. Forordinary end-fire operation the signals from microphones 1, 2, 3, and 4

*An end-fire array has the main directivity lobe along the line of ele-ments as opposed to a broadside array which has the main lobe perpendicularto the line of elements.

(fig. 1) are delayed 0, t, 2t, and 3t seconds, respectively, where t = d/cis the time for sound to travel between microphones.

Microphone arrays are antennas, the general theory of which has beenwell developed (refs. 1-3). Microphone arrays have been used widely inunderwater acoustics (refs. 4-5). Both General Electric Co. (ref. 6) andthe Boeing Co.* have used broadside arrays to measure jet engine noise.Reference 6 has an excellent discussion of the various types of linear arrayswhich could be used specifically for acoustic measurements.

INSTRUMENTATION AND APPARATUS

The array was comprised of four omnidirectional microphones (B&K model4133) mounted to a 2.54 cm (1 in.) diameter pipe as shown in figures 1 and 2.The microphone signals were input to the electronic control module whichdigitized, delayed, summed, and reconverted the signals to analog form. Thedata were then analyzed in third-octave bands. Delay of each channel from0 to 10 milliseconds in 1 microsecond steps could be chosen by the operator.It was discovered that 10 microsecond steps would have been sufficient sincesmaller delay affected signal summation only when the microphones were spacedimpracticably close together. The control module was made entirely with inte-grated circuit components. Design details of the control module will bepublished in a report on the development of an eight channel array.

TESTING AND PROCEDURE

Directivity response was measured in an anechoic chamber by rotating thearray on a turntable 4 m (13 ft) from a fixed loudspeaker. Proper delay forordinary end-fire response depends only on the time of wave propagation fromone element to another, not frequency, and was chosen in such a way that soundwaves traveling along the array summed in phase. The array was also operatedwithout delay resulting in a broadside directivity response.

The array was tested in the 40- by 80-foot wind tunnel at test sectionspeeds of 0, 18, and 28 m/s (0, 58, 92 f/s). The array was aimed at a horndriver mounted near the center of the test section 6.7 m (22 ft) away as shownin figures 2(a) and (b). The omnidirectional microphones with nose cones werepointed into the wind for low wind noise. The source was driven with octavebands of pink (random) noise. The electronic gain of the array output wasremoved from the data so that the array response at a = 00 would have beenequal to that of an omnidirectional microphone with both devices in anechoicspace. The effect of wind speed on wave propagation time was not incorporatedin the time delays used, resulting in some erroneous high frequency array datawhich have been omitted.

RESULTS AND DISCUSSION

Figures 3(a)-(c) show the directivity pattern of the array measured inan anechoic chamber for various values of d/X (ratio of microphone spacingto acoustic wave length). Microphone spacing was 15 cm (0.5 ft). Asexpected, the array performance depended on the ratio d/X. Figure 3(d) showsthat the same directivity resulted for two different spacing (15 cm and 30 cm)

*Conversation with Jack O'Keefe, the Boeing Co.

2

as long as d/A was unchanged. The measured directivity agreed fairly wellwith theoretical predictions based on reference 1 as shown in figures 3(b)and (f). The best directivity resulted when d/X was in the limits of0.35 5 d/X 5 0.88. Above that limit side lobes grew (high frequency) andbelow that limit the main lobe broadened (low frequency). Therefore, if themeasured frequency doubles, the spacing should be changed (for this prototype)to maintain an optimum value of d/X. A nonuniform spacing of 15, 30, and46 cm (0.5, 1.0, 1.5 ft) between successive microphones resulted in the direc-tivity pattern shown in figures 4(a)-(c). The directivity for the nonuniformspacing was no better than that resulting from 30 cm (1 ft) uniform spacing(fig. 3(d)) and has a main lobe with larger side lobes than was measured witha 15 cm (0.5 ft) spacing (fig. 3(b)).

Figures 5(a) and (b) show end-fire response compared to broadside response,the latter response resulting when signal delay was zero. The broadside wasmore directional than end-fire as expected (ref. 6), but the back lobe was asstrong as the front lobe. Since the main lobe of a broadside array is discshaped, sound is accepted equally well from the front, back, above, or below.This is a disadvantage in a reverberant space such as the 40- by 80-footwind tunnel and must be compensated for by using directional microphones ora cross of two arrays which would still have a back lobe (ref. 7). An end-fire lobe, on the other hand, is shaped like a flashlight beam.

Since much aerodynamic noise is broadband and unsteady, array directivitywas measured with the loudspeaker driven by a random noise generator. Figure 6shows that with random noise filtered in octave bands, the valleys betweenminor lobes were eliminated. Since the microphone spacing was fixed, the dif-ference in frequencies between the low and high ends of the octave bandcorrespond to a doubling of d/X. As mentioned above, a doubling of d/Anecessitates a change of microphone spacing. To achieve optimum directivity,therefore, broadband noise should be filtered in third-octave bands ornarrower and spacing should be adjusted for the desired frequency so that d/Ais kept within proper limits. If the directivity of figure 6 is acceptable,octave bands can, of course, be used. The dependency of directivity on fre-quency, and hence d/X, is related to the phase cancellation of off-axissound. The on-axis sound always sums in phase regardless of frequency orfrequency band.

Figures 7(a)-(h) show the results of array acoustic measurements in the40- by 80-foot wind tunnel with and without forward speed compared to theoutput of an omnidirectional microphone. Both receivers had equal sensitivityin the direction of the source (a = 0). The array was successful in reject-ing a substantial amount of reverberation and background noise when aimed ata horn driver in the test section. The source was driven by octave bands ofrandom noise, one band at a time. The source noise at 1000 Hz was partiallymasked by background noise due to wind tunnel operation, but the 2000 and4000 Hz noise was greater than background levels. The data show that thearray rejected 5 to 7 dB of noise in the frequency bands in which the sourcewas operating. In most cases, the array measured the same source noise levelwind on and wind off. Since previous studies (ref. 8) showed that sourcereverberations in the 40- by 80-foot wind tunnel at the array distance are

3

approximately 5 to 7 dB above the direct field, figures 7(a)-(h) indicate thatthe array was measuring free-field noise levels of the horn driver. However,this result was not verified since a free-field calibration of the horn driverwas not made.

The noise rejection of figures 7(a)-(h) is summarized in figures 8(a)-(h)as plots of AdB versus frequency, where

AdB = dB array - dB omnidir. microphone (algebraic subtraction)

The noise outside the frequency bands in which the source was operating wasdue to wind tunnel fan noise and microphone wind noise. The array rejected5 to 12 dB of this noise over a surprisingly large range of d/X(0.18-3.53).This result can be attributed to the fact that 1) on-axis sound sums in phaseregardless of d/X, and 2) side lobes have less effect in a reverberant spacethan figures 3(a)-(i) suggest since it is the response integrated over asphere which determines rejection, not just relative amplitudes of side lobes.Rejection was especially good at frequencies above 400 Hz, but the array wastoo short to effect low frequencies. Longer arrays with more microphonesshould reject noise at lower frequencies. Also, noise rejection by an arraywould be better if directional microphones such as the one described inreference 9 are used instead of omnidirectional microphones.

The amount of microphone noise due to wind in these experiments was notdetermined. However, the relative response of an end-fire to wind-generatedmicrophone turbulence compared to the response to sound waves can be predicted.Figure 9 shows such a response assuming microphone wind noise sums incoher-ently (3 dB for each doubling of the number of microphones) and sound wavesadd coherently (6 dB for each doubling of the number of microphones). Wakeimpingement from one microphone on another, however, should be avoided.

To see if the experimental results of figures 8(a)-(h) are predictable,an analysis was performed to estimate how much noise a four-element arraywill reject in a reverberant space.

Since antenna transmission is analogous to reception (principle of reci-procity), it can be assumed for calculation purposes that the array generatessound. The sound emitted by a source in anechoic space is equivalent interms of amplitude and directivity to the sound received in a reverberantspace. Imagine the emitted sound pressure levels integrated over a spheresurrounding a directional array (i.e., the sound power) in anechoic spacecompared to the sound power of an omnidirectional source adjusted to give thesame pressure level as the array along the array axis (a = 00). Obviously,the omnidirectional source would emit the greater sound power by some amountAdB. Due to the principle of reciprocity, it follows that this AdB soundpower is equal to the AdB difference in sound pressure levels measured byan omnidirectional microphone and a directional array in a reverberant space,both adjusted to have equal sensitivity at a = 00. Thus, the emitted soundpower levels of the array and an omnidirectional source were calculated inthe following manner.

4

The array directivity pattern for 15 cm (1/2 ft) element spacing asmeasured in the anechoic chamber was used to represent transmission direc-tivity. Sound power was calculated using the following relationship (ref. 10):

Lw = Lps + 20 log R + 11 (1)

Lw = total sound power level, dB re 10-12 watt

R = radius of sound measurement locations on a sphere surroundingthe source, meters

Lps = average mean-square sound pressure level over the sphere, dBre 2x10- 5 N/m2.

Each sound pressure measurement was associated with the proper portion of thesphere surface area. It was assumed that the sound pressure levels on acircle around the array were typical of the levels on a sphere around thearray with the array generating sound. The calculated sound power level wasthen subtracted from the sound level of an omnidirectional source whichgenerates the same noise level as the array at a = 00 (i.e., on the arrayaxis).

Figure 10 shows that the calculated results agree fairly well with themeasured values of noise rejection (data from fig. 8(h)). The differencesbetween calculated and measured data were probably due to the wind tunnel testsection being semireverberant rather than reverberant as assumed for thecalculations.

CONCLUDING REMARKS

A four-element, digital-delay, end-fire microphone array was evaluatedfor use as a directional receiver in large wind tunnels. The array directiv-ity measured in an anechoic chamber, though modest, agreed with theoreticalpredictions. The best directivity was attained with the ratio of microphonespacing to acoustic wave length, d/X, kept between 0.35 and 0.88.

Despite the limited directivity of a four-element array, the devicerejected 5 to 12 dB of reverberation and background noise (400-10,000 Hz) inthe Ames 40- by 80-Foot Wind Tunnel. Rejection was good with and withoutforward speed over a range of d/X from 0.18 to 3.53. Limited data indicatedthat the array measured approximately free-field noise levels of a horndriver 6.7 m (22 ft) away. It is estimated that the array will measure 6 dBless wind noise than would an omnidirectional microphone.

Although the array performance in the wind tunnel was much better thanthat of an omnidirectional receiver, longer arrays with more element areneeded to achieve necessary directivity and low frequency discrimination.

5

REFERENCES

1. Kraus, John D.: Antennas. McGraw-Hill, New York, 1950,

2. Collin, Robert E. and Zucker, Francis J.: Antenna Theory, Part I.McGraw-Hill, New York, 1969.

3. Pritchard, R. L.: Optimum Directivity Patterns for Linear Arrays.Acoustics Res. Lab. Tech. Memo. No. 7 (1950), Harvard University.

4. Berktay, H. 0. and Shooter, J. A.: Nearfield Effects In End-Fire LineArrays. J. Acoust. Soc. Amer. 53, 550-563 (1973).

5. Groves, I. D., Jr. and Benedetti, V. P.: Acoustic Measurements with aCircular Nearfied Array in an Anechoic Water-Filled Tank. JASAVol. 54, No. 3, Sept. 1973.

6. Tatge, R. B.: Directive Acoustic Arrays For Jet-Engine Noise SourceLocalization. Report No. 70-C-052, General Electric, Schenectady,New York, Jan. 1970.

7. El-Sum, H. M. A. and Mawardi, 0. K.: Diagnostic Techniques For Measure-ment of Aerodynamic Noise In Free Field and Reverberant Environmentof Wind Tunnels. NASA CR 114636, May 1973.

8. Bies, David A.: Investigation of the Feasibility of Making ModelAcoustic Measurements in the NASA Ames 40- by 80-Foot Wind Tunnel.NASA CR-114352, July 1971.

9. Noiseux, D. U.: Study of Porous Surface Microphones For AcousticMeasurements In Wind Tunnels. NASA CR 114593, April 1973.

10. Beranek, Leo L., ed.: Noise and Vibration Control. McGraw-Hill BookCompany, 1971.

6

Analogoutput

Digitaldelay D/AD/A

A/D ISum

Array axis

Microphone

,-

Wind

Noise source

Figure 1.- End-fire, phased array of microphones.

Altec 806-8Ahorn driver

Lookingupstream

Microphone arrayaimed at source

4.6 m (15 ft)

1.8m (6 ft)

Omnidirectionol 3.7m (12ft) -microphonesfacing into wind

Wind

6.7m (22ft) slant 4.9m (16ft)dist. to array center

Top view oftest section

Test sectioncenterline

(a) Schematic.

Figure 2.- End-fire array and loudspeaker in 40- by 80-foot wind tunnel testsection.

(b) Photograph of array.

Figure 2.- Continued.

(c) Photograph of source.

Figure 2.- Concluded.

250 Hz 15 cm (1/2 ft) spacing

0*

330* 30 °00300

600

dB

2700 40 30 20 10 10 20 30 40900270090°

240 0120 °

2100 1500

1800

(a) d/X = 0.11

Figure 3.- Directivity pattern of four-element array with uniform spacing.Tone source. Directivity angle is relative to array axis.

0 Theory- Experiment


00

330 30

600

dB

2700 40 3 20 10 10 20 30 0900

24001200

210 150

1800

(b) d/X = 0.22



00

3300 300

3000 600

dB40 30 20 1 20 0 30 402700

900

2400 1200

2100 1500

1800

(c) d/A = 0.35


1000 Hz, 15 cm (1/2 ft) spacing--- -- 500 Hz, 30 cm (Ift) spacing

00

330030°

3008 600

dB

27000 20 10 10 20 30 40

24001200

2100 1500

1800

(d) d/. = 0.44



00

3300 30

3000

600

40

2700 30 20 10 10 20 0 409

2400

120 °

2100 1500

1800

(e) d/X = 0.88


0 TheoryExperiment


00

3302 300

dB

270 40 20 10 10 20 40

240"

2100 1500

1800

(f) d/A = 1.33


5000 Hz 15cm (1/2 ft) spacing

330" 300

300 60

dB270040 20 10 1 20 30 40

900

2400 1200

2100 1500

1800

(g) d/X = 2.22



00

3300 30

3000 600

dB

40 30 20 10 O 20 402700 900

2400 1200

2100 1500

1800

(h) d/X = 3.54


10,000 Hz 15 cm (1/2 ft) spacing

00

300 60

dB

27040 10C9 20 40 o0

2400 120

2100 1500

1800

(i) d/X = 4.43


00

3300 300

300" 600

dB

40 30 20 10 10 20 30 402700 900

2400 1200

2100 150"

1800

(a) 500 Hz tone.

Figure 4.- Directivity pattern of array with nonuniform spacing of 15, 30 and46 centimeters (0.5, 1.0, and 1.5 ft), respectively, between microphones.

3302 300

300 60

dB

27040 30 20 10 20 30 490

2400 1200

2100 1500

1800

(b) 800 Hz tone.


00

3302300

3000600

dB

270 40 30 20 1010

24001200

210 0 150 0

1800

(c) 1000 Hz tone.


1000 Hz 15 cm (1/2 ft) spacingd/X = 0.44

00

3300

300 °

600

dB

2700 40 30 20 10 10 20 03 40900

240 0120 °

210 ° 1500

1800

(a) End-fire.

Figure 5.- Comparison of end-fire directivity with broad-side (no delay).Tone source.

1000 Hz, 15 cm (1/2 ft) spacingd/X = 0.44

00

33003 00

3000600

dB

270040 30 20 10 10 20 30 0 40271 i090°

2400 1200

2100 1500

1800

(b) Broad-side.


Source

500 Hz tone 30 cm (1/2 ft) spacing-500 Hz octave band d/X 0.44

00

3300

33000

300 600

dB

2700 40 30 20 10 10 20 30 40 90

2400 1200

2100 - 1500

1800

Figure 6.- Effect of random noise in an octave band.

100

95

Wind tunnel motor noise

90-

NE

85

z

x 80

75 I Iq

*I

A 70- L

L.

60 . I

5- Omnidirectional microphone L---- End-fire array

5020 50 125 315 800 2000 5000 12,50012.5 31.5 80 200 500 1250 3150 8000 20,000

One-third octave bond center frequencies, Hz

(a) Horn off, U = 18 m/s (58 f/s), background noise.

Figure 7.- Comparison of end-fire response with omnidirectional microphoneresponse. Array spacing 15 cm (1/2 ft).

75

- Omnidirectional microphone--- End-fire array

70 -- Background noise (omnidir, mic.)

65 - I

60 -z

0

x 5 5 - --

- 50> II

L-0-45 -

L I

= 40-00o - I

35-

30-

2 5 I I I I I I I20 50 125 315 800 2000 5000 12,500

12.5 31.5 80 200 500 1250 3150 8000 20,000One- third octave band center frequencies, Hz

(b) U = 0, 1000 Hz octave band random (pink) noise from horn.


85-

- Omnidirectional microphone--- End-fire array

80- -o-- Background noise (omnidir. mic.)

75 -

ES70-z

bX 65- L

mImm I

0

60- J

45 -

0

40-I-- ,-

- -,

35 I I I I20 50 125 315 800 2000 5000 12,500

12.5 31.5 80 200 500 1250 3150 8000 20,000One-third octave band center frequencies, Hz

(c) U = 0, 2000 Hz.


75-

-Omnidirectional microphone------ End-fire array

70 - --- Background noise (omnidir. mic.)

65-

ci--E

60 -zz I I

X 55 -

50 -

45 I-- - I) L . I"...., I <I40 - I-- IoI

I L r

35 -LF

II II30

25 I I I I I I20 50 125 315 800 2000 5000 12,50012.5 31.5 80 200 500 1250 3150 8000 20,000

One- third octave band center frequencies, Hz

(d) U = 0, 4000 Hz.


100-

95-

90 - Wind tunnel motor noise

-85-z

x 80-

- I

75 -

I I

70- - L.0 1C L" 65 -

60-Omnidirectional microphone

--- End-fire array55-- Background noise (omnidir. mic.) L-- '-- Background noise (array)

50 I I I I I I20 50 125 315 800 2000 5000 12,50012.5 31.5 80 200 500 1250 3150 8000 • 20,000

One-third octave band center frequencies, Hz

(e) U = 18 m/s (58 f/s), 1000 Hz.


95

90-

85-

E

80

b -

X 75-ej

S70 I

I L

65

: 60 -

SOrnnidirectional microphone- End-fire array--o-- Background noise (omnidir. mic.)

50- - - Background noise (array)

45 1 1 1 1 120 50 125 315 800 2000 5000 12,50012.5 31.5 80 200 500 1250 3150 8000 20,000

One- third octave band center frequencies, Hz

(f) U = 18 m/s (58 f/s), 2000 Hz.


95

90-

85-

E80

x 75-

70 I

65 -

60 -60- I I

-.I

55 - L,- Omnidirectional microphone----- End-fire array

50 --- Background noise (omnidir. mic.)-- o-- Background noise (array)

45 I I I I I I I I I20 50 125 315 800 2000 5000 12,50012.5 31.5 80 200 500 1250 3150 8000 20,000

One- third octave bond center frequencies, Hz

(g) U = 18 m/s (58 f/s), 4000 Hz.


95

- Omnidirectional microphone--- End- fire array

90

85

Ez

80

aox 75-

70-

r J I.

65 - 1 -m _.rC1L

. 60 -0o

55

L50-

45 I I I20 50 125 315 800 2000 5000 12,500


(h) U = 28 m/s (92 f/s), 1000 Hz.


d/X

.04 .06 .09 .14 .22 .35 .55 .89 1.39 2.21 3.53

0

AdB

-10

-2020 50 125 315 800 2000 5000 12,500


(a) Horn off, U = 18 m/s (58 f/s).

Figure 8.- Reduction of reverberant and background noise by the array,

AdB = dBarray - dBomnidir. mic.

d /XI0 .04 .06 .09 .14 .22 .35 .55 .89 1.39 2.21 3.53

0-

AdB

-10

-20 I I I I 1 120 50 125 315 800 2000 5000 12,500


(b) U = 0, 1000 Hz source.


d/x.04 .06 .09 .14 .22 .35 .55 .89 1.39 2.21 3.53

10 I I

0-

AdB

_10

-20

-10 -I

20 50 125 315 800 2000 5000 12,50012.5 31.5 80 200 500 1250 3150 8000 20,000


(c) U = 0, 2000 Hz.


d/X04 .06 .09 .14 .22 .35 .55 .89 1.39 2.21 3.53

10

0-

AdB

-10 -

-20 I I I I I I I I I20 50 125 315 800 2000 5000 12,500

12.5 31.5 80 200 500 1250 3150 8000 20,000One-third octave bond center frequencies, Hz

(d) U = 0, 4000 Hz.


d/X

10 .04 .06 .09 .14 .22 .35 .55 .89 1.39 2.21 3.53

0

AdB

-10

-20 1 I I I I I I I I20 50 125 315 800 2000 5000 12,500


(e) U = 18 m/s (58 f/s), 1000 Hz.


d/X.04 .06 .09 .14 .22 .35 .55 .89 1.39 2.21 3.53

0

0 -

AdB

-10

- 20 I I I I I I I I I I20 50 125 315 800 2000 5000 12,500


(f) U = 18 m/s (58 f/s), 2000 Hz.


d/X04 .06 .09 .14 .22 .35 .55 .89 1.39 2.21 3.53

0-

AdB

-10

-20 L I 1 11 I

20 50 125 315 800 2000 5000 12,50012.5 31.5 80 200 500 1250 3150 8000 20,000


(g) U = 18 m/s (58 f/s), 4000 Hz.


d/X.04 .06 .09 .14 .22 .35 .55 .89 1.39 2.21 3.53

0

AdB

-10

-20 I I I I I I20 50 125 315 800 2000 5000 12,500

12.5 31.5 80 200 500 1250 3150 8000 20,000One-third octave bond center frequencies, Hz

(h) U = 28 m/s (92 f/s), 1000 Hz.


25-

ES20-z

I0

C iOn axis sound from

15 -

0

U)

50

I 2 3 4 5 6 7 8 9 10 1 1 12Number of microphones in array

Figure 9.- Predicted array output in excess of single microphone output.

I0

8

6 0

AdB

4

O Data from figure 8 (h)

2 - Calculated from eqtn. I

I I I I I I I I0 250 500 800 1000 2000 3000 5000 8000 10000

Frequency, Hz

Figure 10.- Calculated and measured noise rejection of array;AdB = dBmdr - dBomnidir. mic. array

nasa technical memorandum€¦ · ordinary end-fire operation the signals from microphones 1, 2, 3,...

Documents